Project summary The proposal ?Accurate measurements of phasing signal to determine macromolecular structures? aims to broaden the applicability of experimental phasing methods, which are an essential component of solving novel structures by X-ray crystallography. The proposed work is significant to the mission of the NIH is due to the importance of X-ray crystallography, which generates uniquely detailed information about cellular processes at the atomic level. This information is used to explain and validate results that are obtain by other biochemical and biophysical techniques, to generate hypotheses for detailed studies of cellular processes by means of orthogonal techniques, and to directly guide drug design studies. Macromolecular crystals are frequently of limited size and crystal lattice order, which leads to decreased quality of phasing signals. Thus, experimentalists average phasing signals across multiple data sets to obtain accurate estimates and to decrease the impact of random errors. The averaging may be performed on data sets acquired from the same crystal but at different wavelengths, different crystals that have been uniformly derivatized with a heavy atom, or data sets collected from different crystals that have not been derivatized uniformly, e.g. different concentrations of heavy atoms were applied during soaking. The averaging of data sets is a challenge because of non-isomorphism between crystals, which may have different impacts on native vs. phasing signals. Additionally, although averaging decreases random errors, the systematic effects in X-ray experiments are not amenable to averaging in the same manner. The optimal approach to diffraction data analysis in the presence of systematic effects is to identify the sources of these effects, and to either eliminate their impact by redesigning the experiment or to filter out their contribution to the phasing signal by applying software corrections. There are several groups of presently uncorrected systematic effects that we have identified and which result in significant distortion of estimates for phasing signals. We propose to develop novel and innovative software corrections to address these problems. In Aim 1, we will develop and implement corrections for systematic inefficiencies present in the signals produced by the current detectors and resulting from sample vibration and complex absorption. In Aim 2, we will develop and implement a novel framework to optimally model the phasing component of the signal obtained by averaging multiple, not-necessarily uniform data sets. In Aim 3, we will develop approaches to filter out uninformative components of the measured intensity, so that the final estimates of the phasing signal are optimal. In Aim 4, a secure, web-based server will be implemented so that the structural community can use our methods, as a complement to the stand-alone software developed in Aims 1-3.